Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
250783 13/6/03 3:34 pm Page 783 JOURNAL OF PLANKTON RESEARCH VOLUME NUMBER PAGES ‒ Vertical microscale patchiness in nano- and microplankton distributions in a stratified estuary LONE THYBO MOURITSEN AND KATHERINE RICHARDSON* DEPARTMENT OF MARINE ECOLOGY, BIOLOGICAL INSTITUTE, AARHUS UNIVERSITY, FINLANDSGADE , DK- AARHUS N, DENMARK *CORRESPONDING AUTHOR: [email protected] Microscale (decimetre) vertical heterogeneity in the distribution of nano- and microplankton was studied on August 23, 1999 at two different sites (separated by ~40 m) in Aarhus Bay, Denmark. At the time of sampling, the water column was stratified with respect to both temperature and salinity and a subsurface fluorescence maximum (corresponding to ~10 µg chlorophyll a l–1) occurred immediately below the primary pycnocline (9 m). Samples for species identification were taken at the surface, near the bottom of the water column and at 20 15-cm intervals in and around the depth of maximum fluorescence. The plankton communities recorded in the three different regions of the water column differed dramatically from one another. In addition, significant differences were found in the distribution patterns of species and functional groups in the region of sampling around the fluorescence peak. The same patterns in vertical species distributions were observed at the two stations. In the region surrounding the subsurface fluorescence peak, diatoms were, generally, regularly distributed, although the total diatom biomass decreased slightly with depth. Dinoflagellate species were mainly non-regularly distributed and could be divided into two groups: (i) autotrophic or potentially mixotrophic species (Dinophysis norvegica, Dinophysis acuminata, Dinophysis cf. dens, Prorocentrum micans, Gymnodinium chlorophorum and Ceratium macroceros) that mainly decreased in numbers with depth or were aggregated in their distribution; and (ii) heterotrophic or potentially mixotrophic species (Ceratium lineatum, Ceratium longipes, Ceratium furca, Ceratium tripos, Ceratium fusus, Protoperidinium curtipes, Protoperidinium steinii, Diplopsalis spp. and Katodinium glaucum). In the latter group, the species mainly increased in numbers with depth or were randomly distributed. Most ciliates were uniformly distributed vertically in the water column. However, small cells of the photosynthesizing ciliate Mesodinium rubrum were most abundant at the depth of maximal fluorescence while large M. rubrum cells were equally abundant at all the depths sampled, suggesting the two size groups of this organism may differ ecologically. Overall, the study demonstrates that traditional plankton sampling methods may lead to misinterpretations of species co-existence and interactions. I N T RO D U C T I O N The patchy distribution in space of plankton organisms under natural conditions is well documented. However, most studies examining this phenomenon have dealt with large scale (i.e. hundreds of metres to kilometres) patchiness in the horizontal plane. Horizontal heterogeneity in plankton distributions is also known to occur at small (i.e. millimetre–centimetre) scales but such patchiness is less well studied and the potential causes of this patchiness less well understood than that occurring at larger scales. Plankton patchiness in the vertical plane is less well studied than that in the horizontal. The most well documented examples of vertical heterogeneity in plankton distributions are the occurrence of subsurface chlorophyll peaks. Such peaks are, most often, identified through fluorescence profiling and are generally considered over scales of metres. Water sampling for the purpose of determining species abundances over the water column are often taken at selected depths but such samples are seldom taken at depth intervals of less than several (often up to tens of ) metres. Thus, although vertical heterogeneity in Journal of Plankton Research 25(7), © Oxford University Press; all rights reserved 250783 13/6/03 3:34 pm Page 784 JOURNAL OF PLANKTON RESEARCH VOLUME phytoplankton distributions is recognized, knowledge concerning the microscale vertical distribution of plankton species is very limited. Regardless of the scale of patchiness one is considering in the ocean, it can come about through either physical (giving rise to heterogeneity in the physical environment) or biological [i.e. migration, sinking, grazing, predation or other more subtle species interactions (Stoecker et al., 1984; Tokarev et al., 1999)] processes or a combination of these two types of process. The importance of biological processes in leading to plankton patchiness is, however, predicted to increase inversely with scale (Pinel-Alloul, 1995). Thus, it is at the microscale where one could predict an important role for biological processes in causing heterogeneity in plankton distributions. Documenting the occurrence of such interactions remains, however, elusive. A few field data sets have provided material on small scale plankton distributions from which individual species’ behaviour or species/community interactions leading to plankton patchiness can be inferred. For example, Bjørnsen and Nielsen (Bjørnsen and Nielsen, 1991) found that high densities of toxic dinoflagellates were associated with a decreased abundance of microzooplankton and suggested a negative interaction between these two plankton groups. In another study, Stoecker et al. (Stoecker et al., 1984) found a positive correlation between the distributions of ciliates and their prey and suggested positive species interactions between these groups. The purpose of the present study was to describe the microscale (decimetre) vertical distributions of phyto- and microzooplankton both at the species and community levels in a stratified estuary. The observed patterns in species distributions are confirmed by sampling at two different stations in close proximity to one another. The study is, in its nature, descriptive. However, we combine our observations on species distributions with previously published information concerning the autecology of the species recorded in order to propose hypotheses to explain how species behaviour or species/community interactions may have influenced the observed patchiness in species and community distributions. METHOD Sampling site Sampling took place in Aarhus Bay, a relatively sheltered estuary situated in the south-western part of the Kattegat, at 56°0910N, 10°1920E, depth 16 m (Figure 1). The hydrographic conditions in the Kattegat are mainly determined by the outflow of low salinity Baltic water at the NUMBER PAGES ‒ Fig. 1. Map of the study area. Arrow denotes location of Aarhus Bay, where the sampling site is located. surface and the inflow of high salinity North Sea and Skagerrak water at the bottom, usually separated by a stable halocline. The vertical position of the halocline is strongly influenced by meteorological conditions such as wind speed and direction as well as atmospheric pressure, which cause the stability of the halocline to vary (Rasmussen, 1997). Moreover, heating of the surface layer by increased irradiance during the summer months reinforces stratification. Sampling Sampling was conducted on August 23, 1999 between 11.00 and 11.30 a.m. using a high resolution sampler (HRS) as described in Bjørnsen and Nielsen (Bjørnsen and Nielsen, 1991). The HRS was positioned so that the middle was placed at the depth of maximal fluorescence (Figure 2). Two HRS deployments were made in the same depth interval at stations separated by a horizontal distance of ~40 m (stations S1 and S2). Prior to HRS sampling, continuous profiles of salinity, temperature, density and fluorescence were recorded in 20-cm intervals in order to determine the depth of the fluorescence peak. CTD measurements were made using an Arop 2000 PowCom and fluorescence measurements of chlorophyll were made with a BackScat fluorometer. Light attenuation in the water column was measured using a Li-Cor quantameter. The HRS collects 20 duplicate water samples (each 13/6/03 3:34 pm Page 785 L. T. MOURITSEN AND K. RICHARDSON VERTICAL DISTRIBUTIONS OF PLANKTON SPECIES microscope (Axiovert 135M, 100, 200, 400). Encountered cells were identified to the lowest possible taxonomic level. When appropriate, cells were assigned to size classes. Nanoflagellates (2–20 µm) were only counted in every second sample (separated by 30 cm). Cells smaller than 2 µm were not counted due to the difficulties of enumerating this size class using the Utermöhl method (Montagnes et al., 1999). Abundance, biovolumes and cell carbon content were calculated using the program Algesys (Bio/consult A/S). Biovolume was calculated by using geometric equations for the shape of the algae (Olrik, 1991). Calculation of cell carbon content followed Edler (Edler, 1979). Only healthy looking cells containing chloroplasts (where appropriate) were enumerated. 0 2 Temperature Salinity 4 Fluorescence 6 Depth (m) 250783 8 10 12 14 16 0 5 10 Temperature (¼C), 15 20 Salinity (ä), 25 30 35 Fluorescence (µg chl/l) Fig. 2. Vertical distribution of temperature, salinity and chlorophyll derived from fluorescence at station S1 on 23 August, 1999. The shaded regions show sampling depths. Twenty depths were sampled (duplicates collected at each depth) using the high resolution sampler (HRS) between 8 and 11 m. Duplicate samples were taken at the surface and in bottom waters. 150 ml) almost simultaneously (<1 s) at 15-cm intervals. From each depth, the two water samples were pooled and a subsample of 50 ml was filtered through a Whatman GF/C filter and frozen for later analysis of nutrient content. The filters were individually placed in 10 ml of 90% acetone for extraction of chlorophyll a. The remaining water was preserved in 1% acidic Lugol’s solution for quantification of phyto- and microzooplankton. In addition to the HRS samples, samples from the surface water (depth = 1 m) and the bottom water (depth = 15 m) were collected and treated as described above. At the time of sampling, wind velocity was low (4–6 m s–1) and there were scattered clouds. Sample analysis Chemical analysis Chlorophyll a concentrations were measured fluorometrically on acetone extracts using a Turner fluorometer, according to Yentsch and Menzel (Yentsch and Menzel, 1963). Analysis of silicate, phosphorus and nitrate concentrations was carried out automatically, following Grasshoff et al. (Grasshoff et al., 1983), at the Danish Institute for Fisheries Research. Taxonomic identification, enumeration and calculation of cell carbon content All diatoms, flagellates and ciliates were quantified by settling 25 or 50 ml of water from each sample for 48 h in plankton sedimentation chambers (Hydrobios). Cells larger than 2 µm were counted following the Utermöhl technique (Lund et al., 1958) using an inverted Zeiss Data analysis Data reduction Various methods have been applied to determine how many cells to count in order to obtain satisfactory precision on cell concentration estimates [e.g. (Allen, 1921; Venrick, 1978)]. In this respect, different authors have reached rather different conclusions, probably as a consequence of, for example, sampling design, the applied counting method and the context in which the data have been used, and/or different authors’ views of ‘satisfactory precision’. Thus, in order to distinguish methodological inaccuracy from real biological patchiness in this study, a separate method, tailored to fit the obtained data and the purpose of sampling, was developed. Species-specific coefficients of variance (CV = variance divided by mean) of cell concentrations from S1 (the 20 samples used as replicates) as a function of mean cell count were used to identify the minimum number of cells counted necessary to stabilize CV. Rejecting the possibility that rare species are inherently more patchily distributed with respect to environmental variables than abundant species, a relatively high CV among the rare species suggests the influence of methodology (i.e. that a few cells by chance appear in one sample whereas none are found in the subsequent sample) rather than the occurrence of real patchiness. This analysis showed that the CV became relatively stable at mean cell counts per sample of around five. One species present in very high numbers in only a few samples (Kinetoplastida sp.) and chain-forming species (Chaetoceros spp. and Skeletonema costatum), which are aggregated in the sedimentation chamber by nature (Holmes and Widrig, 1956), were omitted from the analysis. A comparison of samples from stations S1 and S2 further supported the conclusion that cell counts above five can be considered reliable in the present context. The samples taken in these two HRS deployments demonstrate a remarkable concordance regarding the vertical 250783 13/6/03 3:34 pm Page 786 JOURNAL OF PLANKTON RESEARCH VOLUME distribution of the different species (see Results). Among species with mean cell counts above five, 48% (n = 36) (63% if chain-forming species are omitted) showed a significant species-specific positive correlation (based on Pearson’s correlation coefficients), while only 11% (n = 42) of species with counts of less than five were significantly correlated. In addition, confidence limits, as described in Ricker (Ricker, 1937), were applied to species concentrations recorded in samples from S1 and S2 (upper limit = x + 2.42 + 1.96√[x + 1.5 ]; lower limit = x + 2.42 – 1.96√[x + 0.5]; x = concentration; 95% confidence limits). A number between the concentration limits for each species at each depth was chosen at random and Pearson’s correlation coefficients were calculated. This was done to examine the reproducibility of the number of significant positive correlations found between S1 and S2 applying rather wide confidence limits to the concentration estimates. For species with a high mean number of cells counted (n > 5), the ‘new’ concentrations gave rise to 33% significant positive correlations, which is not statistically different from the original 48% (2 = 0.58, d.f. = 1, P = 0.45). This result indicates that the correspondence of species-specific distribution patterns between S1 and S2 is preserved, even when broad confidence limits are applied to cell concentrations. Based on the above analyses, a mean number of five cells counted was chosen as the reduction parameter, leaving out 36 of the 80 species counted. The following analyses are based on the reduced data set unless otherwise stated. Multivariate analysis Multivariate analyses were performed on biomass data (square root transformed) using the program Primer 4.0 (Clarke, 1993). This was done in order to evaluate possible clustering of samples within and between HRS samples. Using non-metric multidimensional scaling (MDS), which generates a geometric configuration of distances between sample points, derived from a matrix of similarities (Bray–Curtis) between points, two-dimensional scatter plots of the samples were constructed. No fundamental differences were found between MDS plots based on reduced versus unreduced and transformed (fourth and second root) versus non-transformed data, indicating that observed differences between samples are mainly due to the more abundant species. Correlations Spearman’s coefficient of rank correlation (rs) was calculated between species and between nutrients/ salinity/temperature and species in each of the HRS samples. The non-parametric rs is used because a linear NUMBER PAGES ‒ relationship between these parameters cannot be assumed. Species-specific Pearson’s coefficients of correlation (rp) were also calculated between samples from S1 and S2. rp, which assumes a linear relationship between species concentrations, was applied instead of rs in order to emphasize the quantitative relationships between cell concentrations in the spatially separated samples. Assignment to functional groups On the basis of taxonomic and morphological characteristics and a survey of available literature concerning their autecology, the species included in the statistical analyses were assigned to various groups according to (i) taxonomy (diatoms, dinoflagellates, ciliates, other), (ii) size (small, <20 µm; medium, 20–50 µm; large, >50 µm), (iii) mobility (mobile or non-mobile) and (iv) nutritional form (autotrophy, heterotroph or mixotroph). All diatoms were classified as immobile, regardless of reports of migrating diatom mats (Villareal et al., 1993). The classification of species into nutritional groups was the most difficult as the trophic affiliations of several species are currently in question [e.g. (Hansen, 1991a,b; Jacobson and Andersen, 1994)]. The assignments of species to the various groups are summarized in Table I. Distributional patterns of species and groups The vertical distributions of species and groups within the samples taken with the HRS were analysed in different ways. Each species was assigned to a main distribution category: ‘regular’, ‘random’ or ‘aggregated’. Note that the statistical description ‘random’ does not necessarily imply that the distribution of the species in question is incidental or that it is unexplainable but, rather, that it is irregular. The assignment was made by calculating the variance to mean ratio, or index of dispersion (I ) of abundance data using each depth within the HRS as a replicate. I will approximate to unity if there is agreement with a Poisson distribution (that is if data are randomly distributed). The departure of I from unity is assessed by reference to a 2 table, using the approximation 2 = I(n – 1), where n = number of samples and d.f. = n – 1 [see (Elliott, 1977)]. Species that increased or decreased in number with depth were identified by correlating cell concentration and depth using Spearman’s coefficient of correlation. Differences and similarities between distribution patterns of species and groups were analysed using Kendall’s coefficient of concordance (W ) (Sokal and Rohlf, 1995), measuring the degree of association among a number of variables transformed to ranks. W varies between 0 and 1, with W = 1, meaning that there is perfect agreement between the distributions tested, resulting in a significant 2 value (P < 0.05). For each species, 250783 13/6/03 3:34 pm Page 787 L. T. MOURITSEN AND K. RICHARDSON VERTICAL DISTRIBUTIONS OF PLANKTON SPECIES Table I: Feeding form, size and mobility assigned to the different species Dinoflagellates Diatoms Prorocentrum micans (Mi, Me, Mo) Chaetoceros compressus (A, S, Im) Dinophysis acuminata (Mi, Me, Mo) C. curvisetus (A, S, Im) D. cf. dens (Mi, Me, Mo) C. socialis/radians (A, S, Im) D. norvegica (Mi, Me, Mo) C. subtilis (A, S, Im) Ceratium furca (Mi, L, Mo) Guinardia flaccida (A, L, Im) C. fusus (Mi, L, Mo) Leptocylindrus danicus (A, S, Im) C. lineatum (Mi, L, Mo) Guinardia delicatula (A, S, Im) C. longipes (Mi, L, Mo) Rhizosolenia fragilissima (A, S, Im) C. macroceros (Mi, L, Mo) R. pungens (A, L, Im) C. tripos (Mi, L, Mo) Skeletonema costatum (A, S, Im) Gymnodinium chlorophorum (Mi, S, Mo) T. punctigera (A, L, Im) Katodinium glaucum (H, S, Mo) Pseudonitzschia sp. (A, Me, Im) Protoperidinium curtipes (H, L, Mo) Nitzschia closterium/longissima (A, S, Im) P. steinii (H, Me, Mo) Thalassionema nitzschioides (A, S, Im) Diplopsalis spp. (H, Me, Mo) Athecate dinoflagellate 2–10 µm (Mi, S, Mo) Athecate dinoflagellate 10–20 µm (Mi, S, Mo) Thecate dinoflagellate 2–10 µm (Mi, S, Mo) Thecate dinoflagellate 10–20 µm (Mi, S, Mo) Ciliates Others Salpingella sp. (H, Me, Mo) Dichtyocha speculum (A, Me, Mo) Lohmaniella oviformis (H, Me, Mo) Ebria triparthita (H, Me, Mo) Strombidium spp. 20–30 µm (Mi, Me, Mo) Kinetoplastida (Mi, S, Mo) Mesodinium rubrum <25 µm (Mi, S, Mo) Pyramimonas/Tetraselmis (A, S, Mo) M. rubrum >25 µm (Mi, Me, Mo) Prymnesiophyceae spp. (A, S, Mo) Cryptophyceae spp. 14–20 µm (A, S, Mo) A, autotrophic; H, heterotrophic; Mi, mixotrophic; S, small <20 µm; Me, medium 20–50 µm; L, large >50 µm; Mo, mobile; Im, immobile. the relative abundance in each depth (relative frequency of the total number of cells in the given HRS sample) was calculated in order to standardize species distributions. This made it possible to compare species independently of differences in abundance/biomass. Comparisons of vertical distributions of different groups were made by choosing the median frequency of the species in the group (i.e. mobile) in each depth and comparing that with the median frequency of the species in the other group (i.e. immobile). Median frequency was chosen, instead of mean frequency, in order to give weight to typical distributions, thereby minimizing the impact of rare outliers. In the analysis of the agreement of distribution of species with known or potential predator–prey relationships, potential prey organisms were identified from published literature [e.g. (Hansen, 1991; Nielsen and Hansen, 1999; Olli, 1999; Smalley et al., 1999)]. The subsurface fluorescence peak sampled consisted of several minor peaks, all of which were associated with spikes in salinity and temperature. Differences in species composition between the three largest peaks were tested performing Kendall’s coefficient of concordance on the sequence of the 10 most abundant species in terms of biomass. The relative distributions of autotrophic and heterotrophic species in each of these three peaks were examined by comparing the mean of the depth of maximal occurrence in the intervals where the peaks were positioned. Differences were tested for significance using a Student’s t-test. 13/6/03 3:34 pm Page 788 JOURNAL OF PLANKTON RESEARCH VOLUME Statistical analysis NUMBER PAGES ‒ 0 Whenever parametric tests were applied, assumptions of normal distribution and homogeneity of variance were shown to be met using appropriate tests (Kolmogorov–Smirnov and Levenes). Statistical analyses were carried out using the Statistical Package of Social Science (SPSS) unless otherwise stated. 8 P1 Depth (m) 250783 P2 10 P3 R E S U LT S Hydrographic conditions 12 In late August, a primary pycnocline was situated between 5 and 8 m depth. Salinity increased from 19‰ in the surface water to 29‰ at the bottom, while temperatures decreased from 17 to 12°C. Hydrographic conditions were similar at the two stations. Data from station S1 are shown in Figure 2. Salinity and temperature discontinuities usually coincided. Concentrations of nutrients in the depth interval sampled by the HRS were relatively low. The concentrations of phosphate and silicate increased slightly with depth (P: rs = 0.54, n = 20, P = 0.013; Si: rs = 0.61, n = 20, P = 0.004), while the concentration of nitrate was characterized by minor peaks throughout the water column (data not shown). Nutrient concentrations were consistently low in the surface water but higher in the bottom water (phosphate, surface/bottom concentrations: not detectable/1 µmol l–1; nitrate + nitrite: 0.85/0.71 µmol l–1; ammonium: 0.35/4.93 µmol l–1). A fluorescence maximum (corresponding to ~10 µg chlorophyll a 1–1) was measured beneath the primary pycnocline at 9 m depth, corresponding to ~10% of surface photon flux density. Heterotrophs Mixotrophs Autotrophs 16 0 50 100 150 200 250 300 350 µg C/l Fig. 3. Vertical distributions of autotrophic, heterotrophic and mixotrophic biomass (µg carbon 1–1) at the depths sampled at the surface and bottom and with the HRS (station S1). P1, P2 and P3 indicate minor peaks in the autotrophic plankton distribution. Vertical heterogeneity Biomass The fluorescence profile (Figure 2) gives an indication of the overall distribution of autotrophic biomass in the water column and indicates the presence of several distinct peaks in the biomass distribution. The profile of biomass of auto- and mixotrophic phytoplankton estimated by microscopic examination (Figure 3) of the samples in the region of the fluorescence maximum was qualitatively similar to the fluorescence profile, although it appeared that the peaks in biomass were displaced vertically upwards by ~0.5 m between the time of the fluorescence profiling and the water collection by the HRS. Measurements of chlorophyll at the depth interval sampled by the HRS (data not shown) agreed well with the microscopically estimated total auto- and mixotrophic biomass. However, the carbon:chlorophyll ratio decreased significantly with depth (rp = –0.56, P = 0.01), indicating that the phytoplankton in the deeper water layers, where light was less available than at the surface, responded by increasing their cellular chlorophyll content [e.g. (Beardall and Morris, 1976; Richardson et al., 1983; Rosen and Lowe, 1984)]. Species composition Dinoflagellates dominated the plankton biomass between 8 and 11 m. The contributions of ciliates and diatoms to the plankton biomass were lower and relatively constant. A minor increase in mixotrophic biomass was noted in the deepest HRS samples. The dominant species in the samples was Gymnodinium chlorophorum (Elbrächter and Schnepf, 1996), an otherwise colourless dinoflagellate containing an autotrophic endosymbiont. This species was present as colonies embedded in a gelatinous matrix (which is common for this species) and, together with the dinoflagellates Ceratium tripos, Ceratium macroceros, Ceratium fusus, Prorocentrum micans, Dinophysis norvegica and Protoperidinium curtipes, and the diatom Rhizosolenia pungens, comprised ~90% of the total biomass. The dominant species in both the surface and bottom waters were the diatom R. pungens and the dinoflagellate C. tripos. Together, these species comprised 50% of the biomass, while G. chlorophorum (which comprised ~50% of the biomass in the HRS samples) only comprised 14 and 3% in surface and bottom waters, respectively. A closer investigation of the three biomass peaks (P1–P3) studied in the region of maximum fluorescence sampled by the HRS (Figure 3) revealed a similar sequence of the 10 most dominating species (comprising 250783 13/6/03 3:34 pm Page 789 L. T. MOURITSEN AND K. RICHARDSON VERTICAL DISTRIBUTIONS OF PLANKTON SPECIES 90% of the total biomass) in the two uppermost peaks [P1 versus P2: W = 0.96, 29 = 17.2, P = 0.045; NB the subscript on 2 here and in the following represents the number of degrees of freedom in the Chi distribution, see (Sokal and Rohlf, 1995)]. This species sequence was not similar to that of the lowest peak (P1 versus P3: W = 0.81, 29 = 14.8, P = 0.10; P2 versus P3: W = 0.75, 29 = 13.5, P = 0.14). All fluorescence and biomass peaks were associated with changes in salinity and temperature. A difference (although non-significant) was found between the mean vertical positions of autotrophic and heterotrophic species in each of the three peaks, where, in all cases, the autotrophic species peaked above the heterotrophic species. The distances on the MDS plot between the data points representing the plankton communities (unreduced data set) found at the various depths at station S1 indicate a large degree of similarity in the communities recorded in the samples taken with the HRS compared with the communities found at the surface and at the bottom. Likewise, the distance between the points representing the surface and bottom communities on this MDS plot indicates the presence of distinct communities at these depths (Figure 4A). MDS analysis of samples taken with the HRS at both stations (Figure 4B) shows that the shallowest samples cluster at the right of the plot while the deepest are concentrated at the left of the plot. In all cases, samples from the same depths at the two different stations cluster closer together than samples from the top and bottom of the HRS. This indicates a high degree of resemblance for both species composition and biomass at corresponding depths at the two sites, and that the similarity between communities sampled at a specific depth at the two stations (~40 m horizontal distance between stations) was greater than that between communities at the top and bottom of the 3 m sampling range of the HRS at an individual station. Analysis of the species composition showed that a tiny flagellate, Kinetoplastida sp., was present in very high numbers (up to 5500 cells ml–1) but only in the deepest sample taken by the HRS. This observation was made at both stations. Furthermore, species-specific correlations between S1 and S2 showed that 48% of the species were significantly positively correlated (63% if chain-forming species are left out; rp: P < 0.05). However, none of the nanoplanktonic species (<20 µm) was significantly correlated between the two stations. Distribution categories Tables II and III show the results of the statistical distribution analyses. The departure of the index of dispersion (I ) from a Poisson distribution divided the species into three distributional groups (Table II): aggregated (18% of 22 a) 15 10 11 14 13 982456173 18 17 12 19 16 20 21 8 b) 20 16 18 19 38 40 39 36 14 12 6 24 32 26 34 2830 2911 7 25 17 31 35 15 33 9 27 5 3 13 10 23 21 4 22 2 1 Fig. 4. (a) MDS plot of samples taken from all depths at station S1. Calculations are based on the unreduced, square root transformed data set. Samples 1–20 were taken using the HRS in and around the fluorescence peak. Sample 21 is the surface sample and 22 the bottom sample. The plot depicts the uniqueness of the plankton communities occurring at the surface, bottom and middle of the water column. The samples taken in the middle of the water column with the HRS clump together in (a). This clump is expanded in (b): MDS plot of samples taken using the HRS, i.e. only in and around the fluorescence peak at station S1 and S2. Calculations are based on the unreduced, square root transformed data set. Samples from station S1 are numbered 1–20. Samples from corresponding depths at station S2 are numbered from 21 to 40. Points representing samples from corresponding depths at the two stations are connected. Sample 37 was lost. The figure demonstrates both the heterogeneity of the plankton communities sampled at the different depths in the region of the fluorescence peak and the similarity in plankton community structure at the two stations. 250783 13/6/03 3:34 pm Page 790 JOURNAL OF PLANKTON RESEARCH VOLUME NUMBER PAGES ‒ Table II: Distribution patterns of species in HRS samples from station S1 Species Re Prorocentrum micans Ra A * Species Re C. subtilis * Dinophysis acuminata * C. socialis /radians * D. cf. dens * Guinardia flaccida D. norvegica * Leptocylindrus danicus * Ceratium furca * Guinardia delicatula * * Rhizosolenia fragilissima * C. lineatum * R. pungens C. longipes * Skeletonema costatum * * C. macroceros * Thalassiosira punctigera C. tripos * Pseudonitzschia sp. * * * Nitzschia closterium/longissima * Katodinium glaucum * Thalissionema nitzschioides * Protoperidinium steinii * Salpingella sp. * Lohmaniella oviformis * P. curtipes * Diplopsalis spp. * Athecate dinoflagellate 2–10 µm * Strombidium spp. 20–30 µm * Mesodinium rubrum <25 µm * * Athecate dinoflagellate 10–20 µm * Mesodinium rubrum >25 µm Thecate dinoflagellate 2–10 µm * Kinetoplastida sp. Thecate dinoflagellate 10–20 µm * Pyramimonas/Tetraselmis A * C. fusus Gymnodinium chlorophorum Ra * * Dictyocha speculum * Prymnesiophyceae spp. * Chaetoceros compressus * Choanoflagellida spp. * C. curvisetus * Cryptophyceae spp. * Re, regular; Ra, random; A, aggregated. Table III: Species in HRS samples from station S1 whose distributions were found to decrease or increase with depth Species I D Species I Dinophysis acuminata * Dictyocha speculum * D. cf. dens * C. curvisetus * D. norvegica * C. socialis/radians * R. pungens * Ceratium furca * C. fusus * C. macroceros Thalassiosira punctigera * D * Nitzschia closterium/longissima * Protoperidinium steinii * Strombidium spp. 20–30 µm P. curtipes * Mesodinium rubrum <25 µm * * Diplopsalis spp. * Cryptophyceae spp. * I, increasing with depth; D, decreasing with depth. the species), randomly distributed (26%) or regularly distributed (56%). An aggregated distribution was typically ascribed to species with one or more abundance peaks. For randomly distributed species, no pattern in distributions could be discerned and regularly distributed species were more or less uniformly distributed in the region of HRS sampling. The distribution category assigned in Table II describes 3:34 pm Page 791 L. T. MOURITSEN AND K. RICHARDSON VERTICAL DISTRIBUTIONS OF PLANKTON SPECIES the general distribution pattern of the species in the water column. Within all of these categories of distribution, there is the possibility that numbers of an individual species may increase or decrease with depth. Those species for which a significant pattern of distribution with respect to depth in the HRS samples was observed are identified in Table III. The concentrations of 18% of the species decreased significantly with depth, while 24% of the species increased in number with depth (Spearman’s coefficient of correlation, P < 0.05). The remaining species showed no systematic trend with depth. 0 Depth (m) 8 10 Dinophysis norvegica 12 Distribution patterns within taxonomic groups. Within the different taxonomic groups, some patterns in distribution were noted. Almost all of the diatoms encountered in the HRS samples were found to be regularly distributed, although the total diatom biomass decreased slightly with depth. In contrast, the majority of dinoflagellate species were either randomly or aggregatedly distributed in the water column. The dinoflagellates could be divided into two groups with respect to distribution. One group, which consisted of potentially mixotrophic species (G. chlorophorum, D. norvegica, Dinophysis acuminata, Dinophysis cf. dens, P. micans and C. macroceros) showed the same overall distribution pattern as the autotrophic species. Examples of this distribution type are shown in Figures 5 and 6. The other group consisted of heterotrophic species and the remaining mixotrophic species (Ceratium lineatum, Ceratium longipes, Ceratium furca, C. tripos, C. fusus, Protoperidinium curtipes, Protoperidinium steinii, Diplopsalis spp. and Katodinium glaucum). These species mainly increased in 16 0 5 10 15 20 25 30 35 µg C/l Fig. 6. Vertical distribution of the dinoflagellate D. norvegica (µg carbon 1–1) between 8 and 11 m. This distribution pattern is described as aggregated and the cell numbers decrease with depth. Closed circles represent samples taken at station S1, open circles at S2. 0 8 Depth (m) 13/6/03 10 0 8 Protoperidinium steinii 12 16 0 Depth (m) 250783 1 2 3 µg C/l 10 Fig. 7. Vertical distribution of the dinoflagellate P. steinii (µg carbon 1–1) between 8 and 11 m at stations S1 (closed circles) and S2 (open circles). This distribution pattern is classified as regular and increasing with depth. 12 Gymnodinium chlorophorum abundance with depth or were randomly distributed. The distribution of P. steinii is shown in Figure 7 as an example of a typical distribution for this group. Most ciliates were regularly distributed in the water column. 16 0 50 100 150 200 µg C/l Fig. 5. Vertical distribution of the dinoflagellate G. chlorophorum (µg carbon 1–1) between 8 and 11 m at stations S1 (closed circles) and S2 (open circles). This was the most abundant plankter in the HRS samples and its distribution pattern is classified as aggregated. Distribution of groups. No statistically significant similarities in the vertical distribution of various groups of species (grouped according to taxonomy, nutritional form, 250783 13/6/03 3:34 pm Page 792 JOURNAL OF PLANKTON RESEARCH VOLUME mobility and size), were found (Kendall’s coefficient of concordance, dinoflagellates/diatoms/ciliates: W = 0.38, 218 = 20.7, P = 0.29; auto-/mixo-/heterotrophs: W = 0.38, 218 = 15.1, P = 0.66; mobile/immobile: W = 0.66, 218 = 23.8, P = 0.16; small/medium/large cells: W = 0.48, 218 = 25.9, P = 0.10). These tests were performed on the median of the depth-specific frequencies of species comprising the various groups. In contrast, a significant similarity in distribution patterns was found among heterotrophic species (W = 0.34, 219 = 38.3, P = 0.005) and autotrophic species (W = 0.17, 218 = 45.9, P = 0.001), when the concordance of the distribution of the species within these groups was tested. No overall concordance was found among the vertical distributions of species of the other groups. Thus, there appeared to be common factors determining the distribution of species in the autotrophic group and factors influencing the distribution of the species comprising the heterotrophic group. Species–species relationships Of all possible relationships between species, 13% were significantly positive and 7% were significantly negative (Pearson’s coefficient of correlation, P < 0.05). Some of the negative relationships could be explained by major differences in distribution patterns (e.g. increasing or decreasing with depth), suggesting a possible response to a third external factor, for example irradiance or nutrient availability. However, others might be interpreted as being the result of real negative interactions between species. Dinophysis acuminata and D. acuta were, for example, negatively correlated with 7 and 19% of the remaining species, respectively. Of the species negatively correlated with both D. acuminata and D. acuta, no obvious explanation (such as increasing as opposed to decreasing numbers with depth) for the negative associations were found for C. longipes, Rhizosolenia fragilissima, Chaetoceros curvisetus and Leptocylindrus danicus. Nutritional form The distributions of some of the species for which nutritional form is in question or has been demonstrated to vary [e.g. (Hansen, 1991a,b; Jacobson and Andersen, 1994) were examined by calculating Kendall’s coefficient of concordance with respect to the vertical distribution of these species and species representing nutritional groups they could either belong to or be associated with. A high degree of concordance was found between D. norvegica, D. cf. dens and D. acuminata (W = 0.70, 219 = 39.9, P = 0.003). These Dinophysis spp., when tested together with the group ‘heterotrophic dinoflagellates’, gave no concordance (W = 0.08, 219 = 10.4, P = 0.94). However, a significant concordance was found between these Dinophysis spp. and the autotrophic species that decreased with NUMBER PAGES ‒ depth (W = 0.41, 219 = 46.1, P < 0.001). Autotrophs that decreased with depth were tested separately from total autotrophs as the cells that decreased with depth might be expected to be actively photosynthesizing and growing. In contrast, autotrophs that increased in abundance with depth could be sinking or dying cells. Nevertheless, testing Dinophysis spp. together with all autotrophs also gave a significant, although low, concordance (W = 0.15, 219 = 50.5, P < 0.001). Likewise, the distributions of P. micans and C. macroceros were found to resemble the above-mentioned Dinophysis spp. (W = 0.51, 219 = 48.1, P < 0.001) and autotrophic species that decreased with depth (W = 0.31, 219 = 29.1, P < 0.06). No concordance was found between the other Ceratium species (C. furca, C. fusus, C. lineatum, C. longipes, C. tripos) and Dinophysis spp. However, a low but statistically significant concordance (W = 0.21, 219 = 42.7, P = 0.001) existed between the distribution of these Ceratium species and the heterotrophic species (Protoperidinium spp., Diplopsalis spp., K. glaucum, Salpingella sp., Lohmaniella oviformis), suggesting that these species may share a common feeding form. The distribution of G. chlorophorum (Figure 5), which harbours an autotrophic endosymbiont (Elbrächter and Schnepf, 1996), showed a statistically significant similarity with the distribution of Mesodinium rubrum (both size classes combined), which utilizes chloroplasts from ingested cryptomonad prey for photosynthesis (Gustafson et al., 2000), and Strombidium spp. (20–30 µm), which is also possibly mixotrophic (Nielsen and Hansen, 1999) (W = 0.41, 219 = 31.3, P = 0.038). A more detailed analysis of the distribution of M. rubrum cells revealed that while large M. rubrum cells were more or less uniformly distributed, small M. rubrum cells increased with depth in the depth interval surrounding the fluorescence peak (Figure 8). Furthermore, the distribution of the potentially mixotrophic ciliates resembled that of the mixotrophic dinoflagellate G. chlorophorum, which may indicate that feeding mode can be a significant factor influencing the distribution of ciliates. Predator–prey relationships A statistically significant concordance between the distributions of the dinoflagellates Protoperidinium spp. and various potential diatom prey (Chaetoceros spp., Rhizosolenia spp., L. danicus) may suggest a possible predator–prey relationship between Protoperidinium spp. and these species (W = 0.28, 219 = 47.5, P < 0.001). No agreement was found between the distribution of Protoperidinium spp. and the distribution of potential dinoflagellate prey (Dinophysis spp., P. micans, G. chlorophorum, K. glaucum) (W = 0.09, 219 = 15.6, P = 0.68). A similarity in the distribution of the heterotrophic ciliates and potential nanoflagellate prey was found. 13/6/03 3:34 pm Page 793 L. T. MOURITSEN AND K. RICHARDSON VERTICAL DISTRIBUTIONS OF PLANKTON SPECIES indicating that the patterns we observed do not simply reflect a chance distribution of the species recorded at the time of our sampling. A striking demonstration of the similarity in the vertical distribution of species was the fact that one flagellate, Kinetoplastida sp., was found in only one sample at both stations. It was very abundant in both samplings but found only in the deepest sample taken with the HRS. In contrast to the findings for microplankton, the vertical distribution patterns observed for nanoplankton were not found to be similar at both stations. This may suggest that the smaller plankton are not able to control their vertical positions in the water column to the same extent as microzooplankton. 0 < 25 µm 8 > 25 µm Depth (m) 250783 10 Mesodinium rubrum 12 Distribution of diatoms 16 0 1 2 3 4 µg C/l Fig. 8. Vertical distribution in the water column of M. rubrum (cells smaller than 25 µm represented as closed circles; cells > 25 µm as open circles). Both cell sizes classify as being regular in their distribution and the numbers of the smaller cells increase with depth. Significant concordances were found between the ciliate species, Salpingella spp. (W = 0.39, 29 = 28.6, P = 0.001), L. oviformis (W = 0.25, 29 = 17.2, P = 0.038) and Strombidium spp.[(20–30 µm), W = 0.33, 29 = 23.7, P = 0.005] and their potential prey organisms [Cryptophyceae spp. (6–14 µm, 15–20 µm), athecate dinoflagellates (2–10 µm), thecate dinoflagellates (2–10 µm), Pyramimonas/Tetraselmis spp., Choanoflagellida spp. and Chrysochromulina spp.]. A rather high, but non-significant, concordance was found between C. furca and the potential ciliate prey (Strombidium spp. 20–30 µm and Salpingella sp.) (W = 0.45, 219 = 25.8, P = 0.14). DISCUSSION Interpretation of studies of the vertical distribution of plankton species is usually hampered by the time interval occurring between collection of samples at various depths. Here, we have used a novel sampling device that allowed nearly simultaneous sampling at 20 different depths over a 3-m depth range. This allowed us to overcome the problems of time intervals between sampling and turbulence (which potentially disturbs the microscale distributions of plankton species relative to one another) caused by raising and lowering a sampling device in the water column. The study clearly demonstrates microscale vertical heterogeneity in, especially, the distributions of the microplankton species. Furthermore, the major patterns observed in this vertical heterogeneity of microplankton were similar at the two stations studied, Statistical analysis of the distributions of different species and groups of species indicated different patterns of distribution in the region of the water column sampled by the HRS: regular, aggregated or random distribution. Most of the diatoms encountered in the HRS samples were regularly distributed. Exceptions were Guinardia flaccida (randomly distributed), Rhizosolenia pungens (randomly distributed) and Thalassiosira punctigera (aggregated). We have no explanation for why these species might exhibit different general distribution patterns than the rest of the diatoms encountered but note that remnants of an earlier bloom, where cells are now sinking out of the water column, might not be expected to show a regular distribution. Several diatom species were found to decrease in number with depth (Table III), which might be expected to be a response to light availability. Two diatoms, including the aggregatedly distributed T. punctigera, increased in cell number with depth. This, again, might suggest that T. punctigera was sinking out of the water column. Sinking of living and, in particular, dying cells will always be a significant process influencing the distribution of diatoms, especially under conditions of nutrient depletion [(Bienfang et al., 1981) and references therein; (Passow, 1991)]. Some diatom species, (i.e. Chaetoceros spp. and Nitzschia spp.), were mainly found in the upper and lower part of the depth interval sampled by the HRS, i.e. at complementary depths to those where the dominating dinoflagellate G. chlorophorum was most abundant. A possible explanation for this pattern is that the gelatinous matrix of G. chlorophorum may interfere with the spiny structures of Chaetoceros spp. and Nitzschia spp. and, thereby, impede the growth of these species. This hypothesis is supported by the observation that also C. macroceros, a dinoflagellate with very long and slender horns, was also distributed complementarily to G. chlorophorum. Evidence of possible negative interactions between G. chlorophorum and other plankton species has not previously been reported. 250783 13/6/03 3:34 pm Page 794 JOURNAL OF PLANKTON RESEARCH VOLUME In contrast, negative effects of the DSP toxin producing species, D. acuminata and D. acuta, on other microalgae have been demonstrated (Windust et al., 1996). In the present study, D. acuminata and D. acuta were negatively correlated with a number of other species, in particular diatoms, and this can be envisaged to be a result of toxic or inhibitory effects of exudates released from Dinophysis cells. Distribution of dinoflagellates The majority of dinoflagellate species were either randomly or aggregatedly distributed in the region of the HRS sampling. Dinoflagellate distributions could be divided into two groups. One group, which consisted of autotrophic and potentially mixotrophic species, showed the same overall distribution pattern as other autotrophic species (mainly diatoms). The other group consisted of the heterotrophic species and the remaining mixotrophic species. The species in the latter group mainly increased in abundance with depth or were randomly distributed. Examining the first group more closely, we find the most dominant of the organisms sampled by the HRS to be G. chlorophorum. This species was aggregatedly distributed within this depth region and showed greater abundance here than in either the surface or bottom samples. This distribution pattern is unlike that observed by Elbrächter and Schnepf (Elbrächter and Schnepf, 1996), who found G. chlorophorum to accumulate at the surface. However, the bloom in Aarhus Bay, which was first observed at the pycnocline during this study, expanded to cover the upper part of the water column in September and October (P. Andersen, personal communication). Dinophysis norvegica, D. cf. dens and D. acuminata were predominantly found in the upper part of the depth interval, at the level of 10% of surface light, but not in the surface water. A similar observation concerning light preference was made by Carpenter et al. (Carpenter et al., 1995), who found D. norvegica to be most abundant deeper in the water column at low photon flux densities. This may indicate a relatively narrow range of photon flux densities preferred by Dinophysis spp. in contrast to another member of this first group, P. micans, which was uniformly distributed in the HRS samples, abundant in the surface water and rare at the bottom. The wide vertical distribution of P. micans (i.e. the upper two-thirds of the water column) during mid-day has also been demonstrated by other authors (Staker and Bruno, 1980; Eppley et al., 1984; Edler and Olsson, 1985) and suggests a tolerance of a broad span of photon flux densities by this organism. The second group of dinoflagellates, which consisted of the heterotrophic species and the remaining mixotrophic NUMBER PAGES ‒ species, comprised species that mainly increased in abundance with depth or were randomly distributed. Ceratium lineatum and C. longipes were both randomly distributed within the HRS samples and very rare at the surface and the bottom samples. An overall distributional similarity between C. longipes and C. lineatum and the strictly heterotrophic species was demonstrated. No concordance was found in the distribution of C. longipes and potential ciliate prey species. In contrast, the distributions of the other potentially mixotrophic Ceratium species in this group (C. furca, C. tripos and C. fusus) were very wide, with equal concentrations in samples from the surface, the HRS and the bottom, although all three species tended to increase with depth within the HRS samples. The vertical distribution of C. furca around noon has been demonstrated to be wide but also to vary between successive samplings in a number of studies (Blasco, 1978; Eppely et al., 1984; Edler and Olsson, 1985). The distribution of C. fusus has also been found to be wide but dependent on the nutritional state of the cells (Mikaelyan and Zavyalova, 1999). The variable distribution pattern characteristic of C. furca suggests heterotrophy to be a more likely feeding mechanism in this species than autotrophy. Less regularity would be expected in the distribution of heterotrophic cells dispersed in accordance with the dynamic distribution of various prey species than in the distribution of autotrophic cells determined by the less dynamic environmental gradients controlling photosynthesis. A similarity between the distribution of C. furca and of potential ciliate prey was found. Together with a high abundance of ciliates and C. furca, C. fusus and C. tripos at the surface, this suggests the distribution of Ceratium spp. to be strongly influenced by the distribution of prey in the present study. The heterotrophic dinoflagellates P. curtipes and P. steinii were predominantly found in the lower part of the depth interval sampled by the HRS and only rarely in the surface or bottom waters. In contrast, Protoperidinium divergens was more abundant in the surface layer than in the HRS samples. These observations agree well with those of Eppley et al. (Eppley et al., 1984) who demonstrated that P. divergens migrate vertically in the water column and are most abundant in the surface layer at noon. In contrast, Blasco (Blasco, 1978) found another species of the Protoperidinium genus, Protoperidinium depressum, to be distributed predominantly in the subsurface layers, probably as a result of a negative phototactic response. In the present study, the distributions of P. curtipes and P. steinii suggest that the distributions may be influenced by the distribution of potential diatom prey species, indicated by a distributional similarity. The heterotrophic species Diplopsalis spp. and K. glaucum were found mainly in the deepest HRS samples and the 250783 13/6/03 3:34 pm Page 795 L. T. MOURITSEN AND K. RICHARDSON VERTICAL DISTRIBUTIONS OF PLANKTON SPECIES ‘bottom’ sample. Negative phototaxis may explain this distribution but prey availability may also be important. Diplopsalis spp. feed on other large (>20 µm) dinoflagellates (Hansen, 1991b), which were shown to increase with depth in the present study. We might expect the small K. glaucum to feed on nanoflagellates. The concentration of heterotrophic nanoflagellates, which are known to feed on, for example, bacteria associated with sinking, decaying algal cells, can be predicted to increase with depth. In this study, the group Cryptophyceae spp. (14–20 µm), which may include heterotrophic species, was found to increase with depth and the bacteriovore, Kinetoplastida sp., was only found in the deepest samples. This could potentially explain the observed distribution of K. glaucum. Cells of the photosynthesizing ciliate M. rubrum were divided into two distinct size classes, as is customary (Montagnes et al., 1999). Mesodinium rubrum may represent a ‘species complex’ [(Gustafson et al., 2000) and references therein]. Morphological variation is large (Crawford, 1993), as is the variation of cell-specific carbon:chlorophyll ratio and photosynthetic rate (Stoecker et al., 1991). Furthermore, several authors have found the vertical distribution of M. rubrum to vary considerably both on a diel and a seasonal basis (Lindholm and Moerk, 1990; Williams, 1996; Crawford and Lindholm, 1997; Olli, 1999). In the present study, a significant difference was found between the distribution of small (∼15 12 µm) and large (∼25 22 µm) cells of M. rubrum, suggesting the two size classes to be different not only morphologically but possibly also ecologically. The small cells were most abundant in the lower part of the depth interval sampled by the HRS and rare at the surface and bottom, while large cells were uniformly distributed in the HRS samples and abundant in the bottom sample (Figure 8). In a recent study, Pérez et al. (Pérez et al., 2000) also found small M. rubrum cells (length 15 µm) to accumulate around the depth of a subsurface chlorophyll maximum, thus avoiding the surface and bottom layers. Both small and large cells of M. rubrum are believed to engulf Cryptophyceae spp. (Owen et al., 1992; Gustafson et al., 2000). The smallest cryptomonad cells were predominantly found in the lower part of the depth interval sampled by the HRS, coinciding with the small M. rubrum cells, while the larger cryptomonad cells were more randomly distributed, suggesting the distribution of prey cells influences the distribution of M. rubrum. However, nutrient availability (Nielsen and Kiørboe, 1994; Gustafson et al., 2000) and irradiance can also be expected to influence the distribution of photosynthesizing M. rubrum cells. Mobility Two-thirds of the immobile species but only about onequarter of the mobile species were regularly dispersed in the HRS samples (Tables I and II). This pattern, also found by Owen (Owen, 1989) and Olli (Olli, 1999), suggests that the immobile species are subject to dispersal through turbulent water motion, whereas the mobile species are, to a larger extent, able to overcome external physical forcing and exert control over their position within the water column. Size The percentage of species demonstrating a non-regular distribution (i.e. random or aggregated) in the region of the HRS sampling increased with increasing size group. The groups of medium and large species were dominated by mobile hetero- and mixotrophic species, while the group of small species included many immobile autotrophic species but also a number of mobile species. Although the relative swimming velocity of small cells is higher than that of larger cells (Kiørboe, 1993), their ability to actively position themselves in the water column in the face of opposing water motion is smaller. The susceptibility to dispersion by water movements (including those caused by our sampling procedure) may, thus, be expected to be greater in small than in large cells. The influence of hydrography on microscale plankton distributions A comparison of the profiles of fluorescence, nutrient concentrations, salinity and temperature in the depth interval sampled by the HRS revealed a layered structure common to all parameters, suggesting that physical mechanisms also influence vertical microscale plankton patchiness. A multi-layered structure, with discontinuities in salinity and temperature associated with peaks in fluorescence, may result from various hydrographic phenomena, e.g. intrusions of intermediate salinity water transporting nutrients and/or plankton. Differences in the species composition between the lowest and the two uppermost biomass peaks sampled by the HRS also indicate a layered structure consisting of different plankton communities. The overall tendency of auto- and mixotrophic species to peak above the heterotrophic species within the depth interval sampled by the HRS is repeated in each of the minor biomass peaks. The basis for the vertical microscale patchiness observed in the present study may, thus, originate from a hydrodynamically established structure. However, on the basis of existing knowledge concerning the autecology of specific plankton organisms and the statistical analyses reported here, we hypothesize that biological processes (i.e. behaviour of individual species 250783 13/6/03 3:34 pm Page 796 JOURNAL OF PLANKTON RESEARCH VOLUME and/or interactions between species or communities) may influence the fine-scale vertical distribution of individual plankton species. This study indicates that a mid-water column fluorescence peak may, in fact, contain a complex pattern of species distributions that can only be revealed through microscale sampling. We have shown that individual species exhibit distinct and reproducible patterns in their microscale vertical distributions. The presence of vertical microscale heterogeneity as demonstrated here may have important implications for sampling strategies and interpretation of results. Food web structure, species interactions and the relative importance of the species comprising a heterogeneous fluorescence peak may be difficult to infer from a single sample. Furthermore, sampling by vertically positioned water bottles that integrate water from a column the length of the bottle may lead to misinterpretations of species co-existence and species interactions. AC K N O W L E D G E M E N T S NUMBER PAGES ‒ Edler, L. (1979) Recommendations for marine biological studies in the Baltic Sea. The Baltic Marine Biologists Publications, No 5, pp. 1–38. Edler, L. and Olsson, P. (1985) Observations on diel migration of Ceratium furca and Prorocentrum micans in a stratified bay on the Swedish west coast. In Anderson, D. M., White, A. W. and Baden, D. G. (eds), Toxic Dinoflagellates. Elsevier Science Publishing Co., Inc., New York. Elbrächter, M. and Schnepf, E. (1996) Gymnodinium chlorophorum, a new, green, bloom-forming dinoflagellate (Gymnodiniales, Dinophyceae) with a vestigial prasinophyte endosymbiont. Phycologia, 35, 381–393. Elliott, J. M. (1977) Some Methods for the Statistical Analysis of Samples of Benthic Invertebrates, 2nd edn. Scientific Publication No. 25. Freshwater Biological Association, Ambleside. Eppley, R. W., Reid, F. M. H., Cullen, J. J., Winant, C. D. and Steward, E. (1984) Sub-surface patch of a dinoflagellate (Ceratium tripos) off Southern California: patch length, growth rate, associated vertically migrating species. Mar. Biol., 80, 207–214. Grasshoff, K., Ehrhardt, M. and Kremling, K. (1983) Methods of Seawater Analysis. Verlag Chemie, Weinheim, Germany. Gustafson, D. E., Stoecker, D. K., Johnson, M. D., Van Heukelem, W. F. and Sneider, K. (2000) Cryptophyceae algae are robbed of their organelles by the marine ciliate Mesodinium rubrum. Nature, 405, 1049–1052. We would like to thank T. G. Nielsen, National Environmental Research Institute (Denmark) for lending us the HRS. Aarhus County (Denmark) is thanked for allowing us to carry out sampling with their monitoring ship, Thyrfing. This study was partially supported with a grant from the Danish Natural Science Research Council. Hansen, P. J. (1991a) Dinophysis—a planktonic dinoflagellate genus which can act both as a prey and a predator. Mar. Ecol. Prog. Ser., 69, 201–204. REFERENCES Jacobson, D. M. and Andersen, R. A. (1994) The discovery of mixotrophy in photosynthetic species of Dinophysis (Dinophyceae): light and electron microscopial observations of food vacuoles in Dinophysis acuminata, D. norvegica and two heterotrophic dinophysoid dinoflagellates. Phycologia, 33, 97–110. Allen, W. E. (1921) A brief study of the range of error in microenumeration. Trans. Am. Microsc. Soc., 40, 14–25. Beardall, J. and Morris, I. (1976) The concept of light intensity adaptation in marine phytoplankton: some experiments with Phaeodactylum tricornutum. Mar. Biol., 37, 377–387. Hansen, P. J. (1991b) Quantitative importance and trophic role of heterotrophic dinoflagellates in a coastal pelagial food web. Mar. Ecol. Prog. Ser., 73, 253–261. Holmes, R. W. and Widrig, T. M. (1956) The enumeration and collection of marine phytoplankton. J. Cons. Int. Explor., 22, 21–32. Kiørboe, T. (1993) Turbulence, phytoplankton cell size, and the structure of pelagic food webs. Adv. Mar. Biol., 29, 2–72. Bienfang, P., Szyper, J. and Laws, E. (1981) Sinking rate and pigment response to light limitation of a marine diatom: implications to dynamics of chlorophyll maximum layers. Oceanol. Acta, 6, 55–62. Lindholm, T. and Moerk, A. C. (1990) Depth maxima of Mesodinium rubrum (Lohman) Hamburger and Buddenbrock—examples from a stratified Baltic Sea inlet. Sarsia, 75, 53–64. Bjørnsen, P. K. and Nielsen, T. G. (1991) Decimeter scale heterogeneity in the plankton during a pycnocline bloom of Gyrodinium aureolum. Mar. Ecol. Prog. Ser., 73, 263–267. Lund, J. W. G., Kipling, C., Le Cren, E. D. (1958) The inverted microscope method of estimating algal numbers and the statistical basis of estimations by country. Hydrobiologia, 11, 143–171. Blasco, D. (1978) Observations on the diel migration of marine dinoflagellates off the Baja California Coast. Mar. Biol., 46, 41–47. Mikaelyan, A. S. and Zavyalova, T. A. (1999) Vertical distribution of heterotrophic phytoplankton of the Black Sea in summer. Okeanologiya, 39, 895–902 (in Russian). Carpenter, E. J., Janson, S., Boje, R., Pollehne, F. and Chang, J. (1995) The dinoflagellate Dinophysis norvegica: biological and ecological observations in the Baltic Sea. Eur. J. Phycol., 30, 1–9. Clarke, K. R. (1993) Non-parametric multivariate analyses of changes in community structure. Aust. J. Ecol., 18, 117–143. Montagnes, D. J. S., Poulton, A. J. and Shammon, T. M. (1999) Mesoscale, finescale and microscale distribution of micro- and nanoplankton in the Irish Sea with emphasis on ciliates and their prey. Mar. Biol., 134, 167–179. Crawford, D. W. (1993) Some observations on morphological variation in the red-water ciliate Mesodinium rubrum. J. Mar. Biol. Assoc. UK, 73, 975–978. Nielsen, T. G. and Hansen, P. J. (1999) Dyreplankton i danske farvande. (In Danish). Miljø- og Energi- ministeriet. Danmarks Miljøundersøgelser, 1999. Crawford, D. W. and Lindholm, T. (1997) Some observations on vertical distribution and migration of the phototrophic ciliate Mesodinium rubrum (= Myrionecta rubra) in a stratified brackish inlet. Aquat. Microb. Ecol., 13, 267–274. Nielsen, T. G. and Kiørboe, T. (1994) Regulation of zooplankton biomass and production in a temperate, coastal ecosystem. 2. Ciliates. Limnol. Oceanogr., 39, 508–519. Olli, K. (1999) Diel vertical migration of phytoplankton and 250783 13/6/03 3:34 pm Page 797 L. T. MOURITSEN AND K. RICHARDSON VERTICAL DISTRIBUTIONS OF PLANKTON SPECIES Statistics in Biological Research. W.H. Freeman and Co., New York, pp. 593–608. heterotrophic flagellates in the Gulf of Riga. J. Mar. Syst., 23, 145–163. Olrik, K. (1991) Planteplankton metoder. Miljøministeriet, Miljøstyrelsen. Owen, R. W. (1989) Microscale and finescale variations of small plankton in coastal and pelagic environments. J. Mar. Res., 47, 197–240. Owen, R. W., Gianesella-Galvão, S. F. and Kutner, M. B. B. (1992) Discrete, sub-surface layers of the autotrophic ciliate Mesodinium rubrum off Brazil. J. Plankton Res., 14, 97–105. Passow, U. (1991) Species-specific sedimentation and sinking velocities of diatoms. Mar. Biol., 108, 449–455. Pérez, M. T., Dolan, J. R., Vidussi, F. and Fukai, E. (2000) Diel vertical distribution of planktonic ciliates within the surface layer of the NW Mediterranean (May 1995). Deep-Sea Res., 47, 479–503. Pinel-Alloul, B. (1995) Spatial heterogeneity as a multiscale characteristic of zooplankton community. Hydrobiologia, 300/301, 17–42. Rasmussen, B. (1997) The near-surface horizontal buoyancy flux in a highly stratified region, Kattegat. Estuarine Coastal Shelf Sci., 45, 405–414. Richardson, K., Beardall, J. and Raven, J. A. (1983) Adaptation of unicellular algae to irradiance: an analysis of strategies. New Phytol., 93, 157–191. Staker, R. D. and Bruno, S. F. (1980) Diurnal vertical migration in marine plankton. Bot. Mar., 23, 167–172. Stoecker, D. K., Davis, L. H. and Anderson, D. M. (1984) Fine scale spatial correlations between planktonic ciliates and dinoflagellates. J. Plankton Res., 6, 829–842. Stoecker, D. K., Putt, M., Davis, L. H. and Michaels, A. E. (1991) Photosynthesis in Mesodinium rubrum: species-specific measurements and comparison to community rates. Mar. Ecol. Prog. Ser., 73, 245–252. Tiselius, P., Nielsen, G. and Nielsen, T. G. (1994) Microscale patchiness of plankton within a sharp pycnocline. J. Plankton Res., 16, 543–554. Tokarev, Yu. N., Williams, R. and Piontkovski, S. A. (1999) Identification of micro-scale structure of plankton communities of the Black and Ionian Seas by their bioluminescence characteristics. Hydrobiologia, 393, 163–167. Venrick, E. L. (1978) How many cells to count? In Sournia, A. (ed.), Phytoplankton Manual. UNESCO, Paris, pp. 167–180. Villareal, T. A., Altabet, M. A. and Culver-Rymsza, K. (1993) Nitrogen transport by vertically migrating diatom mats in the North Pacific Ocean. Nature, 363, 709–712. Ricker, W. E. (1937) Statistical treatment of sampling processes useful in enumeration of plankton organisms. Arch. Hydrobiol., 31, 68–84. Williams, J. A. (1996) Blooms of Mesodinium rubrum in Southampton Water—do they shape mesozooplankton distribution? J. Plankton Res., 18, 1685–1697. Rosen, B. H. and Lowe, R. L. (1984) Physiological and ultrastructural responses of Cyclotella meneghiniana (Bacillariophyta) to light intensity and nutrient limitation. J. Phycol., 20, 173–183. Windust, A. J., Wright, J. L. C. and McLachlan, J. L. (1996) The effects of diarrhetic shellfish poisoning toxins, okadaic acid and dinophysistoxin-1, on the growth of microalgae. Mar. Biol., 126, 19–25. Smalley, G. W., Coats, D. W. and Adam, E. J. (1999) A new method using fluorescent microspheres to determine grazing on ciliates by the mixotrophic dinoflagellate Ceratium furca. Aquat. Microb. Ecol., 17, 167–179. Yentsch, C. S. and Menzel, D. W. (1963) A method for the determination of phytoplankton chlorophyll and phaeophytin by fluorescence. DeepSea Res., 10, 221–231. Sokal, R. R. and Rohlf, F. J. (1995) Biometry: The Principles and Practice of Received on June 5, 2002; accepted on March 28, 2003